Vol. 69: 261-272, 1991
MARINE ECOLOGY PROGRESS SERIES
Mar. Ecol. Prog. Ser.
Published January 24
Apparent sediment diffusion coefficients for
oxygen and oxygen consumption rates measured
with microelectrodes and bell jars: applications to
oxygen budgets in estuarine intertidal sediments
(Oosterschelde, SW Netherlands)*
P. A. G. Hofman, S. A. de Jong, E. J. Wagenvoort, A. J. J. Sandee
Delta Institute for Hydrobiological Research, Vierstraat 28,4401 EA Yerseke, The Netherlands
ABSTRACT: Oxygen fluxes in shallow marine sediments exceeded calculated oxygen fluxes based on
molecular diffusion of oxygen. Oxygen fluxes measured with bell jars in a stagnant lagoon, combined
with oxygen gradients measured with microelectrodes (bell jar/microelectrode method), resulted in
apparent sediment hffusion coefficients 2 to 3 times higher than the molecular coeff~cient.We have
developed a new method to calculate the apparent sediment diffusion coefficient from oxygen gradient
measurements and independently measured downward-directed oxygen production fluxes, both measured with microelectrodes (microelectrode gradientlproduction method). Diffusion coefficients calculated In this way differed at most 13 % from the apparent sediment diffusion coefficients calculated
using the bell jar/microelectrode method. The microelectrode gradient/production method was applied
to emerged sediments of the Oosterschelde estuary (SW Netherlands), and apparent diffusion coefficients for these sediments exceeded molecular diffusion by 2 to 5 times. Dark oxygen gradients
measured with microelectrodes were used to calculate the oxygen consumption rate of the sediment.
The Q l o value for this rate ranged from 1.3 to 2.7. Two types of oxygen consumption rates were
distinguished. A light-independent oxygen consumption rate depending on diffusion from the overlying
water or air was found, varying between 0.63 to 3.65 mm01 O2 m-2 h-' at different stations. This
consumption increased by 12 to 14 % of the gross oxygen production during the emersion period in
light, in the presence of benthic microalgal production. A net outflux of maximum 45 mm01 O2 m-2 was
found for the entire emersion p e r ~ o dwhereas
,
over a 24 h period production and consumption of oxygen
were in equilibrium. Thus, mineral~zationand respiration processes within the sediment consume most
of the oxygen produced by benthic microalgae on a daily basis.
INTRODUCTION
In recent years great progress has been made in
research on oxygen transport and cycling in marine
sediments, both deep-sea (Reimers et al. 1984, Reimers
1987) and intertidal (Rutgers van der Loeff et al. 1981,
Baillie 1986, Andersen & Helder 1987).
Sediment oxygen consumption can be measured
indirectly in bell jars or cores as the decrease of oxygen
concentration in the enclosed overlying water column
(Rutgers van der Loeff et al. 1981, Van Es 1982, Grant
1986, Reimers & Smith 1986, Andersen & Helder 1987),
Communication no. 508 of the Delta Institute
O Inter-Research/Printed in Germany
or directly in the sediment as described by Revsbech &
Jsrgensen (1983). In bell jars, net oxygen fluxes are
measured. There is no distinction between gross production, consumption and transport of oxygen.
The development of microelectrodes to assess dissolved oxygen in marine sediments (Revsbech et al.
1980) and the light-dark shift method for photosynthetic production measurements (Revsbech & Jsrgensen
1983) are important for s t u d e s of oxygen evolution in
undisturbed sediments. Although microelectrodes precisely measure actual concentrations and thus oxygen
gradients, oxygen fluxes can only b e estimated, since
simultaneous measurement of the sediment diffusion
coefficient is not possible. Oxygen fluxes can be calcu-
Mar. Ecol. Prog. Ser. 69: 261-272, 1991
262
lated with Fick's first law of diffusion: F = - D X 6c/hx,
in which D the diffusion coefficient and 6c/6x = the
slope of the oxygen profile. Jsrgensen & Revsbech
(1985) and Kuenen et al. (1986) calculated oxygen
fluxes from the slope of the oxygen profile in the
diffusive boundary layer (in the overlying water) and
the molecular diffusion coefficient. The apparent sediment diffusion coefficients are often much higher than
molecular diffusion due to bio-irrigation and bioturbation (Krantzberg 1985, Reimers & Smith 1986,
Andersen & Helder 1987, Silverberg et al. 1987),
evapotransport (Baillie 1986), and overlying water
current velocity (Booy et al. in press). Lindeboom et al.
(1985),Reimers & Smith (1986) and Baillie (1986) combined oxygen fluxes from bell jars or incubated cores,
with oxygen gradients in sediments (microelectrode
technique) for apparent diffusion coefficient measurements. They found apparent sediment diffusion coefficients, for shallow-water marine environments, up to
10 times higher than the molecular diffusion coefficient.
For diffusion coefficient and oxygen consumption
rate measurements on emerged sediments these
methods are very inaccurate and do not match in situ
conditions. In bell jars there is oxygen exchange with
overlying water and not with air. Primary production
rates cannot reach the values measured on emerged
sediments due to light extinction (De Jong et al. unpubl.).
In this study we calculated apparent oxygen diffusion coefficients using Fick's first law of diffusion with
(1) oxygen flux measurements in bell jars, combined
with corresponding gradients measured with microelectrodes, and (2) downwards directed oxygen fluxes
-
from direct microelectrode photosynthesis measurements, combined with oxygen gradients. Both methods
were compared in the stagnant saline Lake Grevelingen. The second method was applied on emerged
intertidal flats of the Oosterschelde estuary (SW
Netherlands) to calculate sediment consumption rates.
These consumption rates, combined with oxygen production measurements (De Jong et al. unpubl.), were
used to calculate oxygen budgets over the emersion
period and a 24 h period.
MATERIALS AND METHODS
Investigation area and sampling sites. Oxygen production and consumption were measured at different
locations in the former estuary Lake Grevelingen
(closed in 1971), and in the Oosterschelde estuary, SW
Netherlands (Fig. 1). Lake Grevelingen is now a stagnant lagoon. It is 24 km long, 4 to l 0 km wide and has a
water surface of 108 km2 and a salinity of 30 Ym. During
17 to 22 September 1986, sediment samples were taken
at a water depth of 3 and 7 m (Stn 6 ) . At 3 m, lightlimited benthic primary production took place (De Jong
et al. unpubl.), whereas at 7 m no benthic primary
production could be detected, at light intensities less
than 1% of surface irradiance. The Oosterschelde estuary is a tidal system of ca 370 km2 (45 km length and 4
to 10 km width). In April 1986 measurements were
made at 5 stations chosen on intertidal flats (Fig 1, Stns
1 to 5). Stns 1, 2, 3 and 4 were situated along tidal
channels and were more influenced by tidal dynamics
than was Stn 5, located adjacent to a salt marsh. The
emersion period varied between 5 and 7 h. In general,
Fig. 1. Sampling stahons in the
Oosterschelde Estuary and Lake
Grevelingen, SW Netherlands
Hofman et al.: Oxygen budgets in intertidal sediments
salinity of the interstitial water ranged from 27 to 36 "/00
during the period of emersion. Oxygen consumption
rates and net oxygen production rates were measured
with bell jars and oxygen microelectrodes.
Bell jar method. Bell jars were made of plexiglass
cylinders (i.d.45 cm, height 33 cm). They had a removable lid in which a YSI oxygen electrode (Model no.
5739, Yellow Springs Instr., Inc.), a thermistor and a
stirring device were mounted. The cylinders were
gently pushed into the sediment (3 to 5 cm deep) by
SCUBA divers. The divers checked whether the sediment surface was undisturbed. If so, the lid was closed
and measurement started. Oxygen and temperature
were recorded continuously during 2 to 8 h. Oxygen
production was measured in clear bell jars. For consumption measurements during the light period, some
bell jars were covered with black plastic sheets. The
bell jar measurements were finished before oxygen
concentration reached 50 O/O saturation or 150 % oversaturation in overlying water. Oxygen increase and
decrease were linear. In the light no oxygen bubbles
appeared in the bell jars during the measuring periods
of 2 to 4 h. The water column was stirred continuously
at such a rate that oxygen gradients in the sediment,
measured with microelectrodes inside and outside the
bell jars, showed no significant differences. The macrooxygen electrodes were calibrated a t the beginning
and end of each incubation period, and oxygen concentrations of water overlying the sediment cores were
measured with the Winkler method (Bryan et al. 1976).
At the start of each bell jar experiment, sediment samples were taken directly outside the bell jar with plexiglass corers (i.d. 6 cm), with l 0 to 15 cm overlying water
for microelectrode measurements. At the end of each
experiment, sediment samples were taken outside and
inside the bell jars. Then the bell jars were moved to a
new site, and a new measurement started. Measurement periods lasted 36 h, in which bell jars were moved
5 to 6 times.
For oxygen production and consumption measurements on the emerged flats of the Oosterschelde estuary, 2 cores were taken at the same time and measured
immediately with microelectrodes, one under in situ
light conditions and the other in the dark. A detailed
description of in situ light measurements will be given
in a future paper (De Jong e t al. unpubl.).
Microelectrode method. Primary production and
oxygen micro-gradents were measured with oxygen
microelectrodes as described by Revsbech et al. (1983)
and de Jong et al. (unpubl.). In the construction of
microelectrodes, cellulose nitrate membranes show
several advantages over DePeX a s used by Revsbech &
Jargensen (1983). The cellulose nitrate membrane is
very thin and is simply made by dipping the electrode
in cellulose nitrate dissolved in acetone. For measure-
263
ments, selected microelectrodes had a 90 % response
time <0.05 S and a tip diameter of 5 to 10 pm. Oxygen
profiles were measured with a spatial resolution of
100 pm. Oxygen production by photosynthesizing
benthic microalgae was measured by darkening lightexposed sediment cores for a few seconds and recording the subsequent decrease in oxygen concentration,
as discussed by Revsbech & J ~ r g e n s e n (1983) and
Kuenen et al. (1986). Oxygen production rates were
measured with a spatial resolution of 50 pm (De Jong et
al. unpubl.). Microelectrodes were calibrated by reading the electrode signal in the overlying water above
the boundary layer. The oxygen concentration was
determined by a 2- or 3-fold Winkler titration (Bryan et
al. 1976). The readings in the anoxic part of the sediment were used a s zero values. Electrode readings
were linear with oxygen concentration.
Porosity. Porosity is defined a s the volume of interconnected porewater relative to the volume of total
sediment, provided that it is saturated with water
(Berner 1980, Ullman & AUer 1982). Sediment was
sampled once a day (Oosterschelde intertidal sediment
around low tide), and sliced in 1 or 2 mm layers u p to
10 mm depth. The water content of each 2 mm slice
was related to its porosity. This factor was used to
calculate porosity for the top millimeters of the sediment from the known water content at each sampling
time.
Assumptions a n d calculations.
Assumptions: For diffusion coefficient and oxygen
consumption rate calculations the following assumptions were made:
(1) The oxygen flux into the sediment can b e
described by Fick's first law of diffusion:
where Fo2 = flux of oxygen (mmol O2 m-2 h-'); @ =
mean porosity for aerobic sediment layer; D, = apparent sediment diffusion coefficient (m2h-'); and S02/6x
= oxygen concentration gradient over depth interval X
(mmol O2 m-4).
(2) The apparent sediment diffusion coefficient is
constant over the oxic sediment layer.
Calculation o f the apparent sediment diffusion coefficient: ( A )Bell jar/microelectrode method (Lindeboom
e t al. 1985, Baillie 1986). T h e net oxygen flux (Fo2 be" lar)
across the sediment-water interface is calculated from
the rate of change in oxygen concentration in the water
inside the bell jar. The oxygen gradient (602/6x)in the
sediment is measured with microelectrodes (Fig. 2).
The diffusion coefficient is calculated from:
Da = -Fo2
bell jar/(@
60216~).
P)
Apparent sediment diffusion coefficients, measured
during a dark period with downward oxygen fluxes
Mar. Ecol. Prog. Ser 69: 261-272. 1991
264
SUBMERGED SEDIMENT
a
The apparent diffusion coefficient can be calculated
from:
Oxygen consumption rates on emerged flats can now
be calculated from the dark profiles (Fig. 3c) by combining the value of D,, calculated according to Eq. (3),
with the O2 gradient measured in the dark. For this last
calculation 6O2/8xdarkshould b e constant in time.
RESULTS
Apparent sediment diffusion coefficients in
submerged sedirnents
Bell jar/microelectrode method
I
JJ
<
-g@
Oxygen gradient
N e t oxygen f l u x
Oxygen concentrahon
L~ght
Dark
Fig. 2. Scheme of bell jar/m~croelectrodemethod according to
Lindeboom et al. (1985) (a) Downward oxygen flux (mmol
m-* h-') measured with bell jar, and downward-directed
oxygen gradient (mmol m-4) measured with microelectrode in
the dark. (b)Upward flux and upward-directed gradient in the
light
(Fig. 2a) or during a light period with upward oxygen
fluxes (Fig. 2b), are of the same order of magnitude.
(B) Microelectrode oxygen gradient/production
method. Oxygen production, oxygen consumption and
diffusion of oxygen over depth interval xo - xl result in
an upward-directed oxygen gradient (b02/bxu,) and an
upward-directed flux (Foz Fig. 3a). Oxygen production at depth interval x l - x2,oxygen consumption, and
diffusion of oxygen over depth interval xl -x3 together
result in a downward-directed oxygen gradient (b02/
8xdown).We assumed that the integrated independently
determined oxygen production at depth interval xl -x2
(Foldown) was the source of oxygen over interval x l -x3.
Oxygen
production
Oxygen gradient
in the light
At Stn 6 in Lake Grevelingen, w e measured oxygen
fluxes with bell jars and oxygen gradients with microelectrodes. Porosity was 0.60 ? 0.03 at Site 1, -7 m
(n = 8); and 0.68 k 0.06 at Site 2, -3 m (n = 32). Net
oxygen fluxes measured in bell jars, microelectrodedetermined oxygen gradients and mean porosity of the
upper 2 mm of the sediment were substituted in Fick's
first law of diffusion (Eq. 2), to obtain apparent sediment diffusion coefficients. At the site with a depth of
? m a mean apparent sediment diffusion coefficient of
12.0 X 10P6 m2 h-' (i.e. 3.3 X 1oP5 cm2 S-') (an-' =
1.0 x 10e6, n = 11; 16, l ? , 18 September 1986) was
found; and at the site with a depth of 3 m a coefficient
of 11.8 X 1 0 - ~m2 h-' (anPl = 0.8 X 1 0 - ~ n, = 6; l ? , 18,
19, 22 September 1986) was found (Table 1).
Microelectrode oxygen gradient/production method
Gross oxygen production and oxygen gradients in
the light were measured simultaneously at Stn 6 (Lake
Grevelingen, -3 m) with microelectrodes. Downwarddirected oxygen fluxes, calculated from the production
measurements, and corresponding oxygen gradients
Oxygen gradient
in the dark
R g . 3. Scheme of microelectrode
oxygen
gradient/production
method. (a) Oxygen production
(mM h-') measured with microelectrode, and oxygen fluxes
(mmol m-' h-'). (b) Oxygen gradient (mmol m-4) measured with
microelectrode in the light. (c)
Oxygen g r a h e n t measured in
the dark
Hofman e t al.: Oxygen budgets in intertidal sediments
265
Table 1. Apparent sediment diffusion coefficients ( X 10-%'
h - ' ) , porosity, and production-consumption relation (production:
P,,,,,,; consumption: Ri, mmol m-' h-') for Oostel-schelde Stns 1 to 5 and Grevelingen Stns 6a ( - 3 m) and 6b (-7 m)
Apparent sediment diffusion
coefficient (n)
Stn
Porosity (n)
Production-consumption relation
25.7 ? 5.1" (9)
37.3 ? 8.0" (9)
14.2 k l.?" (5)
12.8 f 3.0" (2)
31.2 k 5.0d (8)
10.3 t I S d (10)
11.8 t 0.8" (6)
12.0 ? l.O"l1)
" Apparent sediment diffusion coefficient calculated by microelectrode oxygen gradient/production method
" Apparent sediment diffusion coefficient calculated by bell jar/microelectrode
were substituted in Eq. (3), and apparent sediment
diffusion coefficients were calculated. We found an
apparent diffusion coefficient of 10.3 X 10-6 m2 h-'
(on_,= 1.5 X I O - ~ , n = 10; 22 September 1986). This
value differs less than 13 % from the sediment diffusion
coefficient found with the bell jar/microelectrode
method.
In a core sampled at a depth of 3 m in Lake Grevelingen, oxygen gradients and gross oxygen production rates were measured at different light intensities,
at time intervals of 10 min (Fig. 4). In this figure the
oxygen gradient (602/6xa,,,,), the corresponding gross
oxygen production and thus the downward-directed
oxygen flux (Fol d o w n ) increased with increasing surface
irradiance. The calculated apparent sediment diffusion
Oxygen ( p M
200
0
600
LOO
method (no production determined)
coefficients were 10.8, 10.5 and 10.8 X 1 0 - ~mZ h-' at
surface irradiances of respectively 83, 172 and 310 p,E
m-2 S-'. Thus, increasing downward-lrected oxygen
fluxes resulted in a corresponding increase in oxygen
gradients, but D, remained constant.
We assumed the downward-directed oxygen production in interval xl - x2 to be an oxygen flux through
interface xl (Fig. 3a). The influence of the known interval thickness (X[ - x2), which differs from the assumed
zero thickness, on the calculation of D, was checked.
We calculated the ratio of X, - x2 (thickness of oxygen
production interval) to X, - x3 (oxygen penetration)
depth), and compared the apparent sediment diffusion
coefficients with this ratio. At stations with a large
number of measurements, no significant correlation
1
Gross oxygen
800
1000
100
0
-1
production p M min )
200
300
LOO
!
I
m
0
-1.
-
0
m Oo
m
0
m
m
E
5
a
X
X
E
O
(U
0
X
0
O o
_
X
X
X
X
P
y
(,
1-
m
eme
o
- o x &
m
X
.*
0
0
l
.
m
e x
80
l
O
&P
* o x ?
0
X
X
xXx
X)*X
2
Fig. 4. Oxygen g r a d ~ e n t sand gross oxygen production measured with microelectrodes in Lake Grevelingen sediment at different
sediment surface irradiances. Measurements were taken at time intervals of 10 min
266
Mar Ecol. Prog. Ser
-
was found at the 5 OO/ level (r = 0.25, n
16; r = 0.46,
n = 11). Thus, D, does not depend on the ratio of
oxygen production interval to oxygen penetration
depth. We conclude that the thickness of the production
interval xl- x2 has no effect on the calculation of D,.
Apparent sediment diffusion coefficients
in emerged sediments
The microelectrode oxygen gradienWproduction
method for calculating apparent sediment diffusion
coefficients was used on the intertidal flats of the Oosterschelde estuary during the emersion period. Dark
oxygen gradients a n d the apparent diffusion coefficients obtained were used to calculate oxygen consumption rate of the sediment (Table 1). The calculated
apparent sediment diffusion coefficients varied from
12.8 X 1 0 - ~m2 h-' (i.e. 3.6 X 10-5 cm2 S-') at Stn 4 to
37.3 X 1 0 - ~
m2 h-' (i.e. 10.4 X 1OP5 cm2 S-') at Stn 2
(April 1986).
Oxygen consumption and production
Net oxygen fluxes and gross oxygen production
in submerged sediments
Measurements of surface irradiance, gross photosynthesis (microelectrodes) and net oxygen fluxes (bell
jars) over a n 11 h period (22 September 1986) in Lake
Grevelingen are shown in Fig. 5. Surface irradiance
reached maximum values of 200 PE m-' S-'. between
13:OO a n d 14:15 h (Fig. 5a). In the underwater sediments of Lake Grevelingen, photosynthesis remained
light-limited. Microalgal biomass was measured, and
apparent microalgal photosynthetic efficiencies were
calculated from incubations (De Jong et al. unpubl.).
Hence, microalgal gross oxygen production rate could
b e calculated a t each light intensity (Fig. 5b). Six time
intervals were distinguished in which the net oxygen
flux measured in the bell jar remained fairly constant
(Fig. 5 c ) . Similar periods were detected in the pattern
of daily surface irradiance. Net oxygen fluxes measured with the bell jars were negative at very low light
intensities at the beginning and end of the light perlod
(minimum value of -0.8 mm01 O2 m-2 h-' ). Net oxygen flux increased up to 3.3 mm01 O2 m-' h-' at the
highest light intensities. For each time interval the
mean net oxygen flux was also calculated from the
microelectrode measurements (gross oxygen production - oxygen consumption). These fluxes corresponded with the bell jar fluxes, although values were
less pronounced and varied between -0.7 and 2.2
mm01 O2 m-2 h-'.
Fig. 5. (a) Irradiance, (b) gross oxygen production calculated
from microelectrode measurements, and (c) net oxygen flux as
measured with bell jars (-)
and microelectrodes ( - - - - -during
)
a light period in Lake Grevelingen (22 Sep 1986) at a depth of
3m
Net oxygen fluxes and gross oxygen production
in emerged sedlments
Gross oxygen production and total oxygen consumption rates were measured with microelectrodes during
the light period on the emerged flats of the Oosterschelde estuary. The results from Stn 6 (Lake Grevelingen) and Stns 1 and 5 (Oosterschelde) are presented in Fig. 6. At the Grevelingen station (22 September 1986) oxygen production started at a light
intensity of about 10 pE m-2 S-' and increased up to 4.4
mm01 O2 m-2 h-' at 200 pE
S-' (Fig. 6a). During the
dark period we calculated a constant oxygen consumption of 1.2 rnrnol O2 m-' h-', using a constant apparent
oxygen diffusion coefficient of 10.3 X I O - ~ m2 h-'.
During the light period, oxygen consumption rates
Hofman et al.: Oxygen bu,dgets in intertidal sediments
267
2.2 mm01 O2 m-' h-' respectively. During the emersion
period oxygen consumption increased to 3.6 and 6.3
mm01 O2 m-' h-' respectively and followed the pattern
of gross oxygen production (Fig. 6b, c).
Oxygen consumption in relation to gross oxygen
production and physico-chemical parameters
rain
0
0
Submerged
Gross production
Consumption
Fig. 6. Gross oxygen production a n d oxygen consumption
measured with microelectrodes during a light period at 3
different stations. (a) Stn 6, in Lake Grevelingen (22 S e p
1986); (b) Stn 1 and (c) Stn 5, both in the Oosterschelde
estuary (25 and 10 Apr 1986 respectively)
showed similar fluctuations to those of gross oxygen
production rates, reaching maximum values of 2.0
mm01 0 2 m-' h-' (Fig. 6a).
Benthic microalgal photosynthesis was limited to the
period of emersion on the intertidal flats of the Oosterschelde estuary. Gross oxygen production started
immediately after the onset of emersion and reached
maximum values of 10.6 mm01 0' m-' h-' at Stn 1
(25 April 1986; Fig. 6b) and 18.0 mm01 O2 m-' h-' at
Stn 5 (10 April 1986; Fig. 6c) (De Jong et al. unpubl.). At
Stns 3 and 5 a steep decrease in production was seen
due to rainfall. During submersion we found a constant
oxygen consumption rate of 1.5 mm01 0' m-' h-' at Stn
1 and 2.8 mm01 O2 m-' h - ' at Stn 5. At the onset of
emersion and the onset of submersion oxygen consumption rates showed a temporary decrease to 1.7 and
Relations between total oxygen consumption rate
and temperature, maximum oxygen saturation and
gross oxygen production rate were calculated. Results
of linear regression (consumption vs maximum oxygen
saturation) and geometric mean regression (consumption vs production) are shown in Fig. 7. The Qlo values
of oxygen consumption rates were 1.3 (at Lake Grevelingen), 1.5 (Oosterschelde Stn 1) and 2.7 (Oosterschelde Stn 5). Good correlations were found between
maximum oxygen saturation and oxygen consumption
(r = 0.92, 0.85, 0.65, and n = 10, 7, 10, respectively).
The correlation coefficients for gross oxygen production vs oxygen consumption rates were 0.66, 0.85 and
0.59 for Stns 6, 1 and 5 respectively (n = 11, 6, 9). In
Table 1 the consumption-production relations are
given for the Oosterschelde intertidal flat stations. An
oxygen consumption rate depending on diffusion alone
and independent of the gross oxygen production rate
was found, ranging from 0.63 to 3.65 mm01 O2 m-2 h-'.
During the light period an additional oxygen consumption rate of 12 to 14 '10 of gross oxygen production rates
was found.
Gross oxygen production and consumption measurements performed in April 1986 were used to calculate
an oxygen budget for all Oosterschelde stations over
the emersion period and a day period (Fig. 8). Stn 2
showed the highest values. The integrated gross oxygen production over the emersion period was 75.8
mm01 O2 m-2, and the oxygen consumption in the light
was 30.6 mm01 O2 m-2, resulting in a net oxygen
outflux during the emersion period of 45.1 mm01 O2
m-'. Over 24 h, a total oxygen consumption of 98.2
mm01 O2 m-' and a net oxygen influx of 22.4 mm01 O2
m-2 were calculated.
Both during the night and during submersion the flux
of oxygen was downward into the sediment. Net oxygen outflux, from the sediment to the atmosphere,
occurred only during the emersion period.
DISCUSSION
Sampling errors
For the micro-gradient measurements with microelectrodes, sediment cores were taken by hand. Under-
Mar. Ecol. Prog. Ser. 69: 261-272, 1991
268
OJ
,
50
100
15
200
Mox~rnol oxygen sotural~oni %l
250
- 2 -1
Fig. 7 Oxygen consumption in
relation to maximal oxygen saturation and gross oxygen production. (a & b) Lake G r e v e h g e n
station (17-22 Sep 1986). (c & d )
Oosterschelde Stn 1 (25 Apr
1986). ( e & f ) Oosterschelde Stn 5
(9, 10 Apr 1986)
Gross oxygen production l mmol O2 m h I
L
m Gross oxygen production
m Oxygen consurnpt~on
Net oxygen influx or outflux
Fig. 8. Gross oxygen production, total oxygen consumption and net oxygen
outflux or influx at 5 intertidal flats in the Oosterschelde
Estuary
(April
1986), calculated over the
light period (L) and over 24 h
(D).Stn l : 25Apr, calculated
with 7 measurements (L)
and 11 measurements (D);
Stn 2: 28 Apr, 6 and 8 measurements respectively; Stn
3: 24 Apr, 4 and 6 measurements; Stn 4 : 23 Apr, 5 and 7
measurements; Stn 5: 9, 10
Apr. 9 and 13 measurements
water sediments can easily b e suspended; thus, sampling has to be done very carefully. Silverberg et al.
(1987) concluded from box-core experiments that little
change in the oxygen gradient occurred during core
retrieval. On the other hand, Reimers (1987) found
discrepancies between microelectrode measurements
in situ a n d those in box-cored deep-sea sediments; she
found that oxygen concentrations in the sediment surface measured in cores were up to twice as high as in
situ values. Our cores were measured within 15 min
after careful sampling, and the oxygen concentrations
in the overlying water did not change during repeated
measurements. The slopes of the oxygen gradients
remained constant for at least 10 min. Hence, apparent
sediment oxygen diffusion coefficients and oxygen
fluxes are not likely to be changed by sampling and
measuring a n d therefore are assumed to b e equal to in
situ values.
Hofman et al.: Oxygen budg ets in intertidal sediments
Determination of apparent sediment oxygen
diffusion coefficients
Bell jar/microelectrode method
For dark oxygen flux measurements in bell jars it is
important that oxygen consumption remains constant
during the measurement. In our experiments, oxygen
could reach concentrations as low as 60 ~ I M
before the
oxygen consumption rate started to decline. A constant
oxygen uptake rate was maintained for several hours in
undisturbed sediment cores where aerobic respiration
predominated (Pamatn~at1971). A nearly constant O2
consumption down to concentrations of 100 yM was
found by Hopkinson (1985) a n d Anderson et al. (1986).
Our bell jar experiments were stopped at half saturahence, oxygen depletion
tion values (ca 150 PM 02);
was not likely to occur.
Stirring of bell jars often does not simulate in situ
water currents (Hopkinson 1985). The flow velocity of
the overlying water influences the oxygen consumption
rate of the enclosed sediment (Bointon et al. 1981, Booy
et al. in press). The benthic boundary layer thickness
and consequently diffusion through this layer depend
on flow velocity (Liss 1973, Broecker & Peng 1974,
J ~ r g e n s e n& Revsbech 1985). Stirring also increases the
penetration depth of the oxygen (Revsbech et al. 1980,
Booy et al. in press). Thus, flow velocities d u e to stirring
in experimental setups have to be checked with in situ
values. Therefore we chose a stirring velocity at which
oxygen gradients measured in sediment inside a n d
outside the bell jars had comparable slopes and depths.
Hence, the apparent sediment diffusion coefficient calculations using Fick's first law, from oxygen flux measurements combined with oxygen gradients (bell jar/
microelectrode method), will be close to in situ values.
Microelectrode oxygen gradient/production method
No correlation was found between the thickness of
the oxygen production layer, responsible for downward-directed oxygen flux, a n d the calculated apparent sediment diffusion coefficient (the 20 % variation in
D, must be the effect of other factors such a s spatial
variation). Different light intensities had no effect on
D,. Thus, considering this layer as an interface should
be a fair approach for the determination of the apparent
sediment diffusion coefficient.
The apparent sediment diffusion coefficients calculated from the downward-directed oxygen production
flux and corresponding oxygen gradients (microelectrode oxygen gradient/production method) showed a
difference of 13 % compared to the bell jar measurements. This difference could be a n effect of infaunal
activity, which does affect the net oxygen flux meas-
269
urements made in bell jars, but it is not included in the
downward-directed production fluxes measured with
microelectrodes. Oxygen gradients can be seriously
disturbed by macrofauna (Meyers e t al. 1987). Only
smoothed, undisturbed oxygen gradients were used for
our calculations, thereby excluding the effects of macrozoobenthos. Meiofaunal oxygen consumption is
included in the microelectrode measurements.
Our calculated apparent sediment diffusion coefflcients (D,) for Grevelingen sediment exceed the molecular oxygen diffusion coefficient (Broecker & Peng 1974)
by a factor of 2. These coefficients are comparable to
those found by Lindeboom et al. (1985), measured at
Lake Grevelingen at a depth of 1.8 m, being 17.1 X I O - ~
m2h-' (calculated as mean sediment diffusion coefficient
a n d corrected for porosity, @ 5 0.7, according to Ullman &
Aller 1982). Revsbech et al. (1980) found apparent
diffusion coefficients ranging between 10.4 x 1OP6 a n d
24.5 X 1oP6 m2 h-' for subtidal sandy sediments.
Increasing diffusion coefficients result in increasing
nutrient fluxes. Silicate fluxes, for example, were found
to exceed molecular diffusion by 2.2 to 6.1 times in
intertidal sediments (Helder & Andersen 1987) and 2 to
10 times in sediments at a depth of 6 to 22 m (Rutgers
van der Loeff et al. 1987), both in the presence of
macrobenthos. Deep-sea sediment research resulted in
oxygen diffusion coefficients near molecular diffusion
(Do),Do = 4.5 X I O - ~ m2 h-' (Reimers & Smith 1986). O n
intertidal flats apparent oxygen diffusion coefficients
1.4 to 3.2 times higher than molecular diffusion (Do =
7.2 x 1 0 - ~m2 h-') were found by Anderson & Helder
(1987). Baillie (1986) calculated apparent oxygen diffusion coefficients ranging from 10.7 to 135.8 X 1OP6 m2
h-'. In the latter 2 studies apparent oxygen diffusion
coefficients were calculated from gradients and fluxes
measured in cores with an overlying water column.
Baillie (1986) immersed her cores taken from exposed
sediment, before measuring oxygen microprofiles.
The apparent sediment diffusion coefficients calculated with the microelectrode oxygen gradient/production method in exposed cores under in situ conditions
fall within in the same range: 12.8 to 37.3 X 1OP6 m2
h-'. However, they are not influenced by manipulation
with overlying water, and approach actual values better. The apparent sediment oxygen diffusion coefficient
is enhanced by the activity of the micro-, meio- a n d
macroinfauna. Their activity results in water currents in
the sediment.
Net oxygen fluxes
The microelectrode production and consumption
measurements were compared with bell jar experiments in Lake Grevelingen (Fig. 5c). Apparent oxygen
270
Mar. Ecol. Prog. Ser. 69: 261-272, 1991
diffusion coefficients measured with both techniques
were similar, apart from the effect of macrofaunal activity, which increased the influx of oxygen to the sediment in the bell jar and thus the calculated apparent
sediment diffusion coefficient.
Net oxygen fluxes measured within the bell jars and
with microelectrodes (gross production corrected for
oxygen consumption in the light) showed some differences but the same trend (Fig. 5c). These differences
can b e explained by phytoplankton production in the
overlying water a t increasing or high light intensities or
by respiration at decreasing or low light intensities. The
bell jar measurements were not corrected for this
phenomenon. Results of oxygen production measurements in the bell jars are very similar to the observations of Lindeboom et al. (1985).
In Lake Grevelingen sediment oxygen production,
and hence production-enhanced oxygen consumption
rates, occurred during the entire light period (Fig. 6a).
On the Oosterschelde intertidal flats, these processes
only took place during emersion in the light (Fig. 6b, c).
As the tide comes in primary production stops, presumably d u e to sudden light extinction and physical disturbance of the sediment-inhabiting microalgae. The water
is very turbid in the Oosterschelde estuary and water
depths increase fast during flood-tide (0.5 m within half
an hour), so that no light reaches the sediment surface.
As a whole, gross oxygen production rates measured
during the emersion period with microelectrodes were
of the same order of magnitude as those found in
intertidal sediments with water-filled bell jars by Rugers
van der Loeff et al. (1981),Van Es (1982), Rizzo & Wetzel
(1985) and Baillie (1986). However, De Jong et al.
(unpubl.) showed that microalgae adapted to a permanent thin water layer exhibited similar chlorophyllbased production rates, but lower gross oxygen production rates, compared to air-exposed microalgae.
At the beginning a n d end of the emersion period,
temporary changes in oxygen gradients were often
noticed. Oxygen gradients were temporarily less steep
and oxygen penetrated deeper into the sediment.
When calculating oxygen fluxes we assumed a constant sediment diffusion coefficient. This resulted in
dips in consumption rates (Fig. 6c). However, it is not
likely that consumption rates change that fast. Probably the incoming or outgoing tide flushes the interstitial water, and hence changes the apparent sediment
oxygen diffusion coefficient temporarily.
Oxygen consumption relations
The calculated Q l o values for oxygen consumption
agree well with literature values, ranging from 1.4 to
3.2 (Pamatmat 1971, Olanczuk-Neyman & Vosjan 1977,
Hopkinson 1985, Andersen & Helder 1987). However,
these Qlo values resulted from field measurements, and
auto-correlations with other factors can play an important role (Hophnson 1985).
Good correlations were found between oxygen consumption and maximum oxygen saturation, and
between oxygen consumption and gross oxygen production. The maximum oxygen saturation is positively
affected by both primary production and temperature.
High oxygen saturation results in oxygen transport to
deeper sediment layers. This results in more oxygenconsuming biomass and oxidation of reduced components in these layers, and hence in higher total oxygen
consumption rates.
Dark oxygen influx of the Oosterschelde intertidal
sediment (0.6 to 3.7 mm01 O2 m-2 h-') IS an order of
magnitude higher than that found in deep-sea sediment [0.02 to 0.15 mm01 m-' h-' (Reimers & Smith
1986, Grant & Schwinghamer 1987, Silverberg et al.
1987)j. Hopkinson (1985) measured oxygen consumption rates of 1.6 to 4.4 mm01 m-2 h-' in his domes,
placed on nearshore sediment without the presence of
benthic primary production. Van Es (1982) found a
dark community respiration on intertidal flats of 0.6 to
2.8 mm01 0' m-' h-' in water-filled bell jars. Our
higher oxygen consumption rates during primary production agree well with those measured by Rizzo &
Wetzel (1985) (3.1 to 4.3 mm01 0' m-' h-') and by
Baillie (1986) (3.5 to 18.7 mm01 m-' h-') on intertidal
flats in the presence of microalgal oxygen production.
Rizzo & Wetzel (1985) also show production-enhanced
oxygen consumption rates on a sandflat and a mudflat
(cf. their Figs. 2 & 3). For phytoplankton, respiration is
assumed to b e proportional to the maximum photosynthetic rate. According to literature cited by Parsons
et al. (1977), the total respiration rate for phytoplankton
is ca 10 '10 of primary production. Our higher values
(dark oxygen consumption increased by 12 to 14 '/o of
gross oxygen production) can be explained by the fact
that microelectrode measurements include all chemical
and biological oxygen-consuming processes. For different stations in the Ems-Dollard Estuary, Van Es (1982)
found a dark oxygen consumption rate of 1.3 mm01 O2
m-2 h-' and a production-enhanced consumption rate
of 38 %. This high factor can be the result of an underestimation of the primary production (measured in
well-miued water-filled bell jars), higher oxygen consumption rates which included macrofaunal respiration, and/or a stirring effect. The above-mentioned
production-enhanced consumption rates all fall within
the range given by Langdon (1988) for phytoplankton
(12 to 40 %).
In the presence of benthic primary producers, up to
between 60 and 70 % of the oxygen produced is consumed immediately within the sediment. This
Hofman et al.: Oxygen budgets in intertidal sediments
phenomenon is described by Kuenen et al. (1986) for
artificial filter biofilms. High oxygen consumption rates
and production-enhanced oxygen consumption rates in
artificial diatom biofilms were found by Jensen &
Revsbech (1989) as well. However, we were able to
measure these processes in natural benthic diatom
mats very accurately with the microelectrode oxygen
gradient/production method. This method permits very
precise calculation of oxygen budgets. Annual oxygen
budgets for the Oosterschelde intertidal sediments will
be published elsewhere (S. A. De Jong & P. A. G.
Hofman unpubl. data).
The sheltered stations (2 and 5) showed the highest
oxygen production and oxygen consumption rates
(Fig. 8), which may be due to a higher organic carbon
content. Nevertheless, on a 24 h basis oxygen consumption equals or exceeds primary production, and no
oxygen is left for non-sediment-inhabiting organisms. A
net oxygen outflux to the water column over a 24 h
period can only be achieved at high microphytobenthos
biomass (visible diatom mats) and high light intensities.
Acknowledgements. This study was part of the BALANSproject granted by Rijkswaterstaat, Tidal Waters Division. We
thank Dr W. Admiraal, Prof. Dr M. Donze, Dr W Helder, Dr
P. Herman, Dr H. Lindeboom, Prof. Dr P. H. Nienhuis, and Dr
H. M Scholten for their critical remarks and fruitful discussions. The crew of the 'Luctor', W. J. L. Rober, J. A. van
Sprundel and P. d e Koeyer, are thanked for their hospitality
and professional help. We are grateful for the assistance of A.
A. Bolsius in preparing the figures for this paper.
LITERATURE CITED
Andersen. F. D., Helder, W. (1987). Comparison of oxygen
microgradients, oxygen flux rates and electron transport
system activity in coastal marine sediments. Mar. Ecol.
h o g . Ser. 37: 259-264
Anderson. L. G., Hall. P. 0 . J., Iverfeldt, A., Rutgers van der
Loeff, M. M-, Sundby, B., Westerlund, S. F. G. (1986).
Benthic respiration measured by total carbonate production. Limnol. Oceanogr 31. 319-329
Baillie, P. W. (1986). Oxygenation of intertidal estuarine sediments by benthic microalgal photosynthesis. Estuar. coast.
Shelf Sci 22: 143-159
Berner, R. A. (1980). Early diagenesis. Princeton University
Press, Princeton, New Jersey
Bointon, W. R., Kemp, W. M , , Osborne, C. G., Kaumeyer, K. R.,
Jenkins, M. C. (1981). Influence of water circulation rate
on in situ measurements of benthic comn~unityrespiration.
Mar. Biol. 65: 185-190
Booy, K., Helder, W., Sundby, B. (in press). Rapid redistribution of oxygen in a sandy sedlment induced by changes in
flow velocity of the overlying water. Neth. J . Sea Res.
Broecker, W. S., Peng, T. H. (1974). Gas exchange rate
between sea and air. Tellus 26: 21-35
Bryan, J . R., Riley, J. P., Williams, P. J. LeB. (1976). A Winkler
procedure for making precise measurements of oxygen
concentrations for productivity and related studies. J . exp.
mar. Biol. Ecol. 21. 191-197
27 1
Grant, J . (1986). Sensitivity of benthic community respiration
and primary production to changes in temperature a n d
light. Mar. Biol. 90 299-306
Grant, J., Schwinghamer, P (1987).Size partitioning of microbial and meiobenthic biomass and respiration on Brown's
Bank, South-West Nova Scotia. Estuar. coast. Shelf Sci. 25:
647-661
Helder, W., Anderson, F. 0. (1981).An experimental approach
to quantify biologically mediated dissolved silicate transport a t the sediment-water interface. Mar Ecol. Prog. Ser
39: 305-3 11
Hopkinson Jr, C. S. (1985). Shallow-water benthic and pelagic
metabolism: evidence of heterotrophy in the nearshore
Georgia Bight. Mar. Biol. 87: 19-32
Jensen, J., Revsbech, N. P. (1989). Photosynthesis and respiration of a diatom biofilm cultured in a new gradient growth
chamber. FEMS Microbial. Ecol. 62: 29-38
Jergensen. B. B.. Revsbech. N. P. (1985). Diffusive boundary
layers a n d the oxygen uptake of sediments and detritus.
Limnol. Oceanogr. 30: 111-122
Krantzberg, G. (1985). The influence of bioturbation on physical, chemical a n d biological parameters in aquatic
environments: a review. Envir. Pollut. 39: 99-122
Kuenen, J . G., Jsrgensen, B. B., Revsbech, N. P. (1986). Oxygen microprofiles of trickling filter biofilms. Wat. Res. 20:
1589-1599
Langdon, C. (1988). O n the causes of interspecific differences
in the growth-irradiance relationship for phytoplankton. 11.
A general review. J . Plankton Res. 10: 1291-1312
Lindeboom, H. J . , Sandee, A. J . J . De Klerk v. d . Driessche,
H. A. J. (1985). A new bell ]ar/microelectrode method to
measure changing oxygen fluxes in illuminated sediments
with a mlcroalgal cover L~mnol.Oceanogr. 30: 693-698
Liss, P. S. (1973).Process of gas exchange across a n air-water
~nterfaceDeep Sea Res. 20: 221-238
Meyers, M B., Fossing, H . , Powell, E. N. (1987). Microdistrlbution of intestitial meiofauna, oxygen and sulfide gradients, and the tubes of macrofauna. Mar. Ecol. Prog. Ser.
35: 223-241
Olanczuk-Neyman, K. M., Vosjan, J. H. (1977). Measuring
respiratory electron-transport-system activity in marine
sediment. Neth. J . Sea Res. 11: 1-13
Pamatmat, M. M. (1971). Oxygen consumption by the seabed.
IV Shipboard and laboratory experiments. Limnol.
Oceanogr 16: 536-550
Parsons, T R., Takahashi. M - , Hargrave, B. (1977). Biological
oceanographic processes. Pergamon Press, Oxford
Reimers, C. E. (1987). An in situ microprofiling instrument for
measuring interfacial pore water gradients: methods a n d
oxygen profiles from the North Pacific Ocean. Deep S e a
Res. 34: 2019-2037
Reimers, C. E., Kalhorn, S.. Emerson, S. R., Nealson, K. H.
(1984). Oxygen consumption rates in pelagic sediments
from the Central Pacific: first estimates from microelectrode profiles. Geochim. cosmochim. Acta 48: 903-910
Reimers, C. E., Smith, K. L. J r (1986). Reconciling measured
and predited fluxes of oxygen across the d e e p sea sediment-water interface. Limnol. Oceanogr 31: 305-318
Revsbech, N. P., Jsrgensen, B. B. (1983). Photosynthesis of
benthic microflora measured with high spatial resolution
by the oxygen microprofile method: capabilities and limitations of the method. Limnol. Oceanogr. 28: 749-756
Revsbech, N. P., Jsrgensen, B. B., Blackburn, T H., Cohen, Y .
(1983). Microelectrode studies of the photosynthesis a n d
02, H2S and pH profiles of a microbial mat. Limnol.
Oceanogr. 28: 1062-1074
Revsbech, N. P , Ssrensen, J . , Blackburn, T H., Lomholt, J . P.
272
Mar. Ecol. Prog. Ser. 69: 261-272, 1991
(1980). Distribution of oxygen in marine sediments measured with rnicroelectrodes. Limnol. Oceanogr. 25: 403-411
Rizzo, W. M , , Wetzel. R. L. (1985).Intertidal and shoal benthic
community metabolism in a temperate estuary: studies of
spatial and temporal scales of variability. Estuaries 8:
342-351
Rutgers van der Loeff, M. M . , Anderson, L. G., Hall, P. 0. J . ,
Iverfeldt, A., Josefson, A. B., Sundby, B., Westerlund,
S. F. G , (1984). The asphyxiation technique: an approach
to distinguishing between molecular diffusion and biologically mediated transport a t the sediment-water interface.
Limnol. Oceanogr. 29: 675-686
Rutgers van der Loeff, M. M . , Van Es. F. B., Helder, W.. De
Vries, R. T. P. (1981). Sediment water exchanges of
nutrients and oxygen on tidal flats in the Ems-Dollard
Estuary. Neth J . Sea Res. 15: 113-129
Silverberg, N., Bakker, J., Edenborn, H. M., Sundby, B. (1987).
Oxygen profiles and organic carbon fluxes in Laurentian
Trough sediments. Neth. J . Sea Res. 21: 95-105
Ullman, W. J., Aller, R. C. (1982) Diffusion coefficients in
nearshore marine sediments. Lirnnol. Oceanogr. 27:
552-556
Van Es, F. B. (1982). Community metabolism of intertidal flats
in the Ems-Dollard Estuary. Mar Biol. 66: 95-108
This article was submitted to the editor
Manuscript first received: June 14, 1989
Revised version accepted: November 2, 1990